SEC Investigating Magnetar, JP Morgan Dealings on Subprime CDO

ProPublica reports that the SEC has taken interest in Magnetar’s role in a JP Morgan underwritten CDO.

We discussed Magntar at length in our book ECONNED, which broke that story six weeks before ProPublica launched its report, and remains the definitive account of how those transactions were structured. (Our continuing beef with the ProPublica account is that it missed what we discussed at length in ECONNED: the systemic impact of the Magnetar trade. It isn’t simply that Magnetar was a bad actor; its Constellation CDO program played a direct and substantial role in increasing the severity and damage of the toxic phase of the subprime bubble).

Per ProPublica’s update:

The Securities and Exchange Commission is investigating whether JPMorgan Chase allowed a hedge fund to improperly select assets for a $1.1 billion deal backed by subprime mortgages, according to people familiar with the probe.

Called “Squared” and completed in May 2007, the deal was a collateralized debt obligation, or CDO, made up of pieces of other CDOs. The hedge fund, Magnetar Capital, based in Evanston, Ill., purchased the riskiest slice of Squared as part of a strategy to bet against the mortgage market.

The issue is similar to the one that came up in the SEC suit against Goldman in April over a 2007 Abacus trade (this was one deal in a much larger Goldman program called Abacus): did subprime short John Paulson, who did take a short position in that trade, act as a Trojan horse long for a small percentage of the deal (the equity tranche) so as to gain influence over what bonds went into the deal? With the Paulson involvement, it was harder to argue impropriety, since he had made the fact that he was shorting subprime public, and the CDO manager, ACA, (who was nominally responsible for picking the exposures and clearly was negotiating with Paulson what was in and out of the deal) was part of one of the major investors in the long side of the deal (in other words, if one hand didn’t know what the other at ACA was doing, you could hardly blame the failure to communicate on Paulson and Goldman).

The trick was that firms like Magnetar and Paulson were sponsors of synthetic or heavily synthetic CDOs (note in the Abacus trade, weirdly, Paulson merely acted as if he was legitimately at the table negotiating the exposures;, remarkably, he didn’t act as the deal sponsor). The equity tranche was normally the most difficult to place, and in return for taking that risk, the sponsor typically got the right to influence the deal, in theory to reduce the risk for all investors. The minimum right was being able to nix particular exposures, but as the Paulson/Abacus example indicates, some investors went further and actually presented lists of desired assets (previous reports on Magnetar in the Wall Street Journal also indicate that Magnetar selected particular bonds). But the equity position was a sham; both Paulson and Magnetar took short positions well in excess of their equity tranche position, making them net short..

As we noted in ECONNED:

Anyone involved in these transactions probably understood the implicit logic, even if no one acknowledged it. But there is a remarkable absence of anyone who could be pinned with liability. Magnetar officially had no legal relationship to these deals. The investment bank packager/structurer was off the hook as long as he made reasonable disclosure (and remember, the standards are much lower here than for instruments that fall in the SEC’s purview). The rating agencies get off scot-free, thanks to their First Amendment exemption (discussed in chapter 6). The lawyers involved in the deal are responsible only to their clients, meaning the structurer/packager, and cannot be sued by unhappy investors. The only party on whom liability could be pinned is the CDO manager, who does have a fiduciary responsibility to all investors, not just the sponsor. But the fact that the party who in theory had the most to lose, Magnetar, approved their investments, would seem to exculpate the CDO manager.

The ruse of diffusing responsibility, and of having the party most clearly liable, the CDO manager, be an economically weak party (CDO managers typically were very small shops, sometimes with as few as a couple of professional staff), means these cases are hard to prove, even if the nature of the chicanery seems obvious now.

The JP Morgan transaction was called “Squared” because it was a CDO squared, meaning a CDO made from (typically) the riskier tranches of unsold CDOs, with a bit of other types of credit exposures thrown into the mix to make it look slightly less unsavory. The rather peculiar thing about Squared is it consisted heavily of exposures from other Magnetar CDOs, which strongly suggests Magnetar had a hand in this deal (as in dealers were choking on the unsold exposures, and Magnetar needed them placed to launch new first gen CDOs).

But Mr. Market does seem to be taking this announcement seriously; the averages started moving down when this story broke today, and Bloomberg attributes the decline to this suit. Although JP Morgan was much less heavily involved in synthetic and heavily synthetic CDOs than Goldman, if the SEC makes a case that sticks against a Magnetar trade, every major firm in the subprime business will be at risk. This could get interesting, in a good way for a change.